thinking about a 200 amp 12v pwm controller

[izzi]

I"ve started looking into putting together around a 200-250 amp 12v pwm
controller for a DC drive i have. I've got the pwm driving aspect of this
down aside from the mosfet selection and dealing with the heatsinks that
i'll be stuck using. Anyone ever put something along these lines together
before? The cycle time will be low around 1 min on and very long time off.
30min+ range, this should help with heat issues. Any input is welcome.


** Posted from http://www.teranews.com **

 

[Jon Slaughter]

"izzi" <nope@test.com> wrote in message
news:51f91$48e1a450$26355@news.teranews.com...

I"ve started looking into putting together around a 200-250 amp 12v pwm
controller for a DC drive i have. I've got the pwm driving aspect of this
down aside from the mosfet selection and dealing with the heatsinks that
i'll be stuck using. Anyone ever put something along these lines together
before? The cycle time will be low around 1 min on and very long time off.
30min+ range, this should help with heat issues. Any input is welcome.

** Posted from http://www.teranews.com **

No, but you can easily get mosfets within that range(I think anyways).. You
are looking for a very low rdson and I think it means your going to have to
parallel mosfets. (theres an article on the net on how to safely do this)

I have some 40V 80A 2.5mOhm 300W power fets that if you paralleled n of them
you could have

n*80 Amps,
2.5/n mOhms.

Hence the power dissipation per fet is

(I/n)^2*2.5/n = 0.0025*I^2/n^3

With I = 200, n = 4, you have only 1.5W per mosfet.

and total amperage is 4*80 = 320A so you have a little headroom to
spare(might need to up it to 5)

Basically 5 of these mosfets in parallel act like one 50V 400A 0.5mOhm
mosfet.

Of course this doesn't take into account switching losses which make the
results invalid. (But formula would be similar)

Point being that paralleling mosfets is like paralleling resistors... it
reduces the current on each one but allows you to increase the total
current. Of course if one goes out then you end up increasing the current
through the remaining which could end up putting to much through them.
you'll need to search for paralleling mosfets to see how to do it properly
and probably need some sort of saftey mechanisms(maybe somehow use a
multiple relay for each branch that will cut all branches if one isn't
conducting due to a bad mosfet).

In any case I'd probably try to find properly rated mosfets first

 

[Nemo]

I"ve started looking into putting together around a 200-250 amp 12v pwm
controller for a DC drive i have. I've got the pwm driving aspect of this
down aside from the mosfet selection and dealing with the heatsinks that
i'll be stuck using. Anyone ever put something along these lines together
before? The cycle time will be low around 1 min on and very long time off.
30min+ range, this should help with heat issues.

Concerning the heatsink, you may want to look into heat pipes:

http://en.wikipedia.org/wiki/Heat_pipe#External_links

basically these are "super heat sinks". A bit of googling will find you
various suppliers. In the past I have heard of custom ones being used
for cost-no-object apps - but there may well be some standard ones
around you can get fairly cheaply. There are restrictions on orientation
and, I think, ambient temperature.

Of course the best option is to avoid generating heat in the first
place, you want every last % of efficiency, which can be helped by
things not normally relevant to modern micro designs like using
larger-than-minimum-size inductors (lower losses).
--
Nemo